Autonomous Agent Harness
affaan-m/ECC
Transform Claude Code into a fully autonomous agent system with persistent memory, scheduled operations, computer use, and task queuing.
Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both.
MITAuto-check passed
Install Loop Reconcile
$ npx skills add rome-os/rome --skill loop-reconcile -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install rome-os/rome loop-reconcile --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/loop-reconcile .claude/skills/loop-reconcile && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .claude/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcileType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add rome-os/rome --skill loop-reconcile -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install rome-os/rome loop-reconcile --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/loop-reconcile .agents/skills/loop-reconcile && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .agents/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add rome-os/rome --skill loop-reconcile -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install rome-os/rome loop-reconcile --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/loop-reconcile .cursor/skills/loop-reconcile && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .cursor/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/rome-os/rome.git --path .claude/skills/loop-reconcile--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add rome-os/rome --skill loop-reconcile -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install rome-os/rome loop-reconcile --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/loop-reconcile .gemini/skills/loop-reconcile && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .gemini/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install rome-os/rome loop-reconcileInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add rome-os/rome --skill loop-reconcile -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/loop-reconcile .github/skills/loop-reconcile && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .github/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add rome-os/rome --skill loop-reconcile -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install rome-os/rome loop-reconcile --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/rome-os/rome.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/loop-reconcile .opencode/skills/loop-reconcile && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "loop-reconcile" agent skill from https://github.com/rome-os/rome/tree/main/.claude/skills/loop-reconcile into .opencode/skills/loop-reconcile/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "loop-reconcile", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
loop-reconcile- GitHub stars
- 725
- Token cost
- ~1.6k tokens
- SKILL.md length
- 949 words
- Files
- 1
- Skills in repo
- 18
- Repo updated
- First seen
- Licence
- MIT
At a glance
Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both.
- Works in 3 steps: Guard. Read the open issues carrying… → Compute the gap. Walk the Statements… → Exit by what the walk found. A run files…
- SKILL.md covers Inputs and precedence, Authority, Procedure and Open questions, plus 1 more section
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Loop Reconcile is an agent skill from rome-os/rome. Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both.
Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
The repository describes itself as: A compounding agent OS for recursive agents. Also an open source alternative to Grok Bot and Meta's Muse. The licence is MIT.
Example prompts
- “/loop-reconcile”
Workflow steps
3 steps, taken from the first numbered list in SKILL.md.
- Guard. Read the open issues carrying loop-reconcile and this run's scope label.
- Compute the gap. Walk the Statements list one claim at a time against the code. Record each as holds, with evidence, or fails, with the…
- Exit by what the walk found. A run files at most one report.
What it can do on your machine
Read from SKILL.md and the folder at commit ed26d88. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Loop Reconcile loads about 1.6k tokens when it runs. Until then it costs about 41 tokens; SKILL.md has 949 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from rome-os/rome at commit ed26d88, republished under its MIT licence (© rome-os). 949 words, ~1,556 tokens.
Download SKILL.mdSave it as
.claude/skills/loop-reconcile/SKILL.md (or your agent's skills folder).- name
- loop-reconcile
- description
- Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both.
- disable-model-invocation
- true
Loop: reconcile
Goal end-state: every deliverable chunk the run can scope exists as issues, awaiting ready-for-agent, and one report explains whatever it could not scope.
This is the autonomous half of the reconciliation loop. When launched from a session, run it in a background agent. It runs headless: never ask the user a question mid-run. Every ambiguity resolves to a report, never to something plausible. Every report reads standalone on GitHub and never depends on chat context.
All work state lives in the GitHub repository itself: issues, labels, and milestones. No project board. Every issue this skill files carries two labels: loop-reconcile, the mark that the loop owns it, and a scope label loop:<doc-stem> naming its source doc — loop:people for docs/northstars/people.md. A report is an issue too, and it carries loop-complete or loop-needs-discussion on top of those two. If a label is missing from the repository, create it.
Args: an ideal-state doc under docs/northstars/, and optionally a repository milestone.
- Given a milestone, reconcile against the milestone description. First check that the doc still backs every milestone statement: if one contradicts the doc, or the doc no longer asks for it at all, file a
loop-needs-discussionreport and stop. The doc wins on intent, and a milestone that fell behind it needs a session, not work. - Given no milestone, reconcile against the doc's Statements section and file issues attached to no milestone.
- Only the Statements list has authority. The rest of the doc is illustration.
Inputs and precedence
| Source | Role | Authority |
|---|---|---|
| Ideal-state doc | What should be | Wins on intent |
| Codebase | What is | Wins on facts |
| Issue tracker | What is in flight | None — it gates and remembers |
- The tracker never influences what to scope. It only gates whether to scope.
- Do not read PR or git history to guess intent. If a gap looks like a deliberate move away from the doc, scope it anyway. The human withholds
ready-for-agentand fixes the doc. - Concurrency belongs to the driver. Assume this run is alone.
Authority
- Allowed: create issues, comment, and revise or withdraw issues that carry
loop-reconcileplus this run's scope label and notready-for-agent. - Never: apply
ready-for-agent, create or advance milestones, close an issue labeledready-for-agent, touch an issue outside this run's scope label, or edit the ideal-state doc. - Anything a human approved is frozen. Issues labeled
ready-for-agentand milestones change only by human hand.
Procedure
- Guard. Read the open issues carrying
loop-reconcileand this run's scope label.- If one is labeled
ready-for-agentor is in progress, file nothing. Print a pointer to it and stop. - If none carries
ready-for-agent, re-check each against the gap computed below. If they hold, stop. If they are stale, revise or withdraw them, then continue.
- If one is labeled
- Compute the gap. Walk the Statements list one claim at a time against the code. Record each as holds, with evidence, or fails, with the shortest path to make it hold.
- Exit by what the walk found. A run files at most one report.
- A statement cannot be checked, or two statements cannot both hold: the doc is broken. File a
loop-needs-discussionreport naming the statements, and file nothing else. The doc gets fixed in aloop-northstarsession, never by this run. - Zero gap: file a
loop-completereport walking each statement with its evidence and naming what it checked — the doc, or the milestone. Never advance to the next milestone. - Otherwise: scope chunks by the rules below and file them as issues, up to about four. Alongside them file a
loop-needs-discussionreport for every question the walk could not answer and every chunk the cap left out. Chunks and a report are one run's ordinary output, not alternatives.
- A statement cannot be checked, or two statements cannot both hold: the doc is broken. File a
The size of a gap decides how much of it a run files. It never decides whether the run files anything.
Show full SKILL.md (318 more words)Show less
Open questions
A report names the statements it questions. Every other statement the walk covered is unquestioned, and a chunk serving only unquestioned statements is filable while that report is open.
- The report enumerates the live resolutions of each question it raises. A chunk is filable when every one of those resolutions still needs it. A chunk that some resolutions need and others do not belongs to the report.
- A filable chunk touches no file the contested fix would touch. This is the mechanical check when the resolutions are hard to enumerate.
- A filable chunk names the report, states the open question in one line, and states which resolutions it survives. That claim is what a human checks before applying
ready-for-agent. - Its Scope names the contested area as out of scope, so the boundary is written down rather than inferred.
- A report questioning a statement carries the replacement text it proposes. A
loop-northstarsession then reviews a draft rather than redesigning from the problem.
Scoping rules
- A chunk fits one reviewable PR.
- Change a shared contract by expand and contract: add the new alongside the old, and move consumers in later chunks. If the contract and all its consumers fit one chunk, change it in place.
- The chunk that removes the last consumer of a replaced thing also deletes it. Dead code is noise to the next gap computation.
- When an issue states a fact about the code, it names the file that shows it. The reader verifies by looking, not by searching. This caps nothing — a chunk touches as many files as one reviewable PR allows.
- File several issues only when the chunks are parallel: disjoint files, no shared contract. Never file a dependency chain — a chained issue is a stored prediction that goes stale.
- Issues follow the issue format, attach to the milestone when one is given, and stand alone without the doc.
- Never apply
ready-for-agent.
© rome-os, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in .claude/skills/loop-reconcile of rome-os/rome.
Open the folder on GitHubat commit ed26d88
Compare with similar skills
Loop Reconcile next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Loop Reconcile this skillrome-os/rome | 725 | — | ~1.6k | Automated safety check: Pass | MIT | |
| Autonomous Agent Harnessaffaan-m/ECC | 275k | 2 repos | ~2.9k | Automated safety check: Pass | MIT | |
| Autonomous Loopsaffaan-m/ECC | 275k | 4 repos | ~5.8k | Automated safety check: Pass | MIT | |
| Phoenix Docs Gap AuditArize-ai/phoenix | 12k | — | ~4.8k | Automated safety check: Pass | Custom licence | |
| Docnexu-io/open-design | 100k | — | ~266 | Automated safety check: Pass | Apache-2.0 | |
| Docsremotion-dev/remotion | 62k | — | ~248 | Automated safety check: Pass | Custom licence |
Similar skills
Autonomous Loops
affaan-m/ECC
Patterns and architectures for autonomous Claude Code loops — from simple sequential pipelines to RFC-driven multi-agent DAG systems.
Phoenix Docs Gap Audit
Arize-ai/phoenix
Audit documentation gaps across the Phoenix repo by analyzing recent commits to main (default: last 7 days).
Doc
nexu-io/open-design
Read, create, and edit .docx documents with formatting and layout fidelity via OpenAI's document skill.
- Official
Docs
remotion-dev/remotion
Start the Remotion docs site and open it in the Codex browser.
- Official
Update Docs
vercel/next.js
This skill should be used when the user asks to "update documentation for my changes", "check docs for this PR", "what docs need updating", "sync docs with code", "scaffold docs for this feature"…
More from rome-os/rome
All 18 skills in this repoColor Audit
rome-os/rome
Audit a design system's color palette against measurable color-science disciplines — WCAG/APCA contrast of declared token pairs, perceptual (OKLCH) ramp uniformity, color-blindness safety of…
Secondary Text Audit
rome-os/rome
Audit the subordinate copy in a UI — section descriptions, field helper text, hints, card subtitles, empty-state body copy, tooltip bodies — against the secondary-text ruleset, and emit a per-string…
UX Semantics Audit
rome-os/rome
Audit an existing UI/UX design (React/JSX/TSX components, HTML, or generated app code) against a tiered ruleset of verifiable UX principles, and produce structured, evidence-cited findings that a…
Add OAuth Integration
rome-os/rome
Add a new Rome-managed OAuth integration for a third-party service so a user can delegate access by clicking Connect, and Rome can act on the service with the delegated token (the GitHub/Slack model…
Composio CLI
rome-os/rome
Help users operate the published Composio CLI to find the right tool, connect accounts, inspect schemas, execute tools, subscribe to trigger events with composio listen, script workflows with…
File Issue
rome-os/rome
File one GitHub issue from a description the user gives — classify it as a bug report, feature request, or task spec, gather what the body needs from the tracker and the code, ask the user only for…
Questions about Loop Reconcile
What does Loop Reconcile do?
Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both. Loop Reconcile is an agent skill from rome-os/rome. Autonomously compute the gap between an ideal-state doc and the codebase, then file the next work issues, a report on what needs discussion, or both.
How do I install Loop Reconcile in Claude Code?
Run `npx skills add rome-os/rome --skill loop-reconcile -a claude-code`. Or copy the skill folder (.claude/skills/loop-reconcile in rome-os/rome) into .claude/skills/loop-reconcile in your project. Claude Code loads it when a task matches its description.
How do I install Loop Reconcile in Codex?
Run `npx skills add rome-os/rome --skill loop-reconcile -a codex`. Or copy the skill folder (.claude/skills/loop-reconcile in rome-os/rome) into .agents/skills/loop-reconcile in your project. Codex loads it when a task matches its description.
Can I use Loop Reconcile in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add rome-os/rome --skill loop-reconcile -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/loop-reconcile, .gemini/skills/loop-reconcile, .github/skills/loop-reconcile and .opencode/skills/loop-reconcile in your project.
What does Loop Reconcile need to run?
SKILL.md names no scripts, command-line tools or credentials: Loop Reconcile is instructions for the agent only.
Does Loop Reconcile access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Loop Reconcile safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Loop Reconcile use?
Loop Reconcile is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Loop Reconcile use?
About 1.6k tokens (SKILL.md is roughly 6.2k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Loop Reconcile?
Skills that share tags, products or a category with Loop Reconcile: Autonomous Agent Harness (affaan-m/ECC, 275k stars), Autonomous Loops (affaan-m/ECC, 275k stars), Phoenix Docs Gap Audit (Arize-ai/phoenix, 12k stars) and Doc (nexu-io/open-design, 100k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Loop Reconcile?
rome-os (a GitHub organization) maintains it in rome-os/rome, which has 725 GitHub stars. The repository holds 18 skills in this directory. The repository was last updated on October 8, 2026.
Source: rome-os/rome on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.